ValueError: arrays must be the same length

[peter-ch]

I'm running the demo notebook unmodified on Ubuntu 20.04 machine with one GPU, everything is pip installed properly, but I keep getting this error below no matter what I do. It works in Colab but not locally. Why? How can I fix this?

| load data: .//StockTradingEnv_trading_data.df
| YahooDownloader: finish processing date list
| GPU id: 0, cwd: ./AgentPPO/StockTradingEnv-v2_0
Loaded act: ./AgentPPO/StockTradingEnv-v2_0

ValueError Traceback (most recent call last)
in
6 args.init_before_training()
7
----> 8 prediction = args.env.trade_prediction(args, torch)

~/Dropbox/Code/work/stocks_boredhumans/Elegant-FinRL/elegant_finrl/env.py in trade_prediction(self, args, _torch)
266 # print(len(self.date_list))
267 # print(self.date_list)
--> 268 df_account_value = pd.DataFrame({'date':self.date_list,'account_value':episode_returns})
269 return df_account_value
270

/usr/local/lib/python3.8/dist-packages/pandas/core/frame.py in init(self, data, index, columns, dtype, copy)
527
528 elif isinstance(data, dict):
--> 529 mgr = init_dict(data, index, columns, dtype=dtype)
530 elif isinstance(data, ma.MaskedArray):
531 import numpy.ma.mrecords as mrecords

/usr/local/lib/python3.8/dist-packages/pandas/core/internals/construction.py in init_dict(data, index, columns, dtype)
285 arr if not is_datetime64tz_dtype(arr) else arr.copy() for arr in arrays
286 ]
--> 287 return arrays_to_mgr(arrays, data_names, index, columns, dtype=dtype)
288
289

/usr/local/lib/python3.8/dist-packages/pandas/core/internals/construction.py in arrays_to_mgr(arrays, arr_names, index, columns, dtype, verify_integrity)
78 # figure out the index, if necessary
79 if index is None:
---> 80 index = extract_index(arrays)
81 else:
82 index = ensure_index(index)

/usr/local/lib/python3.8/dist-packages/pandas/core/internals/construction.py in extract_index(data)
399 lengths = list(set(raw_lengths))
400 if len(lengths) > 1:
--> 401 raise ValueError("arrays must all be same length")
402
403 if have_dicts:

ValueError: arrays must all be same length

[jhmenke]

For a temporary fix you can use self.date_list[1:] since the first entry is wrong. The Yahoo Downloader can return dates previous to the start date.

A better fix is to add two lines after the first one in env.py:

print("| YahooDownloader: finish processing date list"). # this one exists already
raw_df["date"] = pd.to_datetime(raw_df.date)
raw_df = raw_df[raw_df.date >= datetime.strptime(start, "%Y-%m-%d")]

[SteamedGit]

I'm going to turn this into a pull request. Credit to @jhmenke.